CN107356556A - A kind of double integrated modelling approach of Near-Infrared Spectra for Quantitative Analysis - Google Patents
A kind of double integrated modelling approach of Near-Infrared Spectra for Quantitative Analysis Download PDFInfo
- Publication number
- CN107356556A CN107356556A CN201710569312.8A CN201710569312A CN107356556A CN 107356556 A CN107356556 A CN 107356556A CN 201710569312 A CN201710569312 A CN 201710569312A CN 107356556 A CN107356556 A CN 107356556A
- Authority
- CN
- China
- Prior art keywords
- sample
- training
- subset
- quantitative analysis
- submodel
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000002329 infrared spectrum Methods 0.000 title claims abstract description 18
- 238000004445 quantitative analysis Methods 0.000 title claims abstract description 13
- 238000013459 approach Methods 0.000 title claims abstract description 11
- 238000012549 training Methods 0.000 claims abstract description 58
- 238000000034 method Methods 0.000 claims abstract description 39
- 241000131894 Lampyris noctiluca Species 0.000 claims abstract description 18
- 238000012614 Monte-Carlo sampling Methods 0.000 claims abstract description 6
- 238000005457 optimization Methods 0.000 claims description 4
- 238000007796 conventional method Methods 0.000 claims description 2
- 239000000126 substance Substances 0.000 abstract description 6
- 239000003814 drug Substances 0.000 abstract description 2
- 238000002835 absorbance Methods 0.000 description 21
- 239000000203 mixture Substances 0.000 description 11
- 238000010586 diagram Methods 0.000 description 8
- KFZMGEQAYNKOFK-UHFFFAOYSA-N Isopropanol Chemical compound CC(C)O KFZMGEQAYNKOFK-UHFFFAOYSA-N 0.000 description 6
- LFQSCWFLJHTTHZ-UHFFFAOYSA-N Ethanol Chemical compound CCO LFQSCWFLJHTTHZ-UHFFFAOYSA-N 0.000 description 4
- 238000002790 cross-validation Methods 0.000 description 4
- 239000000446 fuel Substances 0.000 description 4
- 238000005259 measurement Methods 0.000 description 4
- 238000001228 spectrum Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 3
- 239000000295 fuel oil Substances 0.000 description 3
- 230000003595 spectral effect Effects 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 238000000465 moulding Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 2
- 241000254158 Lampyridae Species 0.000 description 1
- 238000010521 absorption reaction Methods 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000008033 biological extinction Effects 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000012417 linear regression Methods 0.000 description 1
- 239000003921 oil Substances 0.000 description 1
- 238000010238 partial least squares regression Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/17—Systems in which incident light is modified in accordance with the properties of the material investigated
- G01N21/25—Colour; Spectral properties, i.e. comparison of effect of material on the light at two or more different wavelengths or wavelength bands
- G01N21/31—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry
- G01N21/35—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light
- G01N21/359—Investigating relative effect of material at wavelengths characteristic of specific elements or molecules, e.g. atomic absorption spectrometry using infrared light using near infrared light
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/004—Artificial life, i.e. computing arrangements simulating life
- G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
Landscapes
- Physics & Mathematics (AREA)
- Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Spectroscopy & Molecular Physics (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Theoretical Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Pathology (AREA)
- Biochemistry (AREA)
- Analytical Chemistry (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Immunology (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Chemical & Material Sciences (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
The present invention relates to a kind of double integrated modelling approach of Near-Infrared Spectra for Quantitative Analysis, comprise the following steps that：The sample of certain amount is collected, the near infrared spectrum of sample is gathered, takes certain packet mode to split data into training set and forecast set；The sample for choosing certain amount from training set by the use of Monte Carlo sampling method is used as training subset；Using glowworm swarm algorithm, further selected part variable as sample variable subset, establishes partial least square model from the training subset；Successive ignition establishes some models；Final prediction result is obtained to the result progress simple average of all models.The present invention uses double integrated modelling approach, improves the precision of prediction and stability of Near-Infrared Spectra for Quantitative Analysis.The present invention is applied to the quantitative analysis field of the complex samples such as chemical industry, Chinese medicine and food.
Description
Technical field
The invention belongs to the Chemical Measurement modeling technique of analytical chemistry field, and in particular to a kind of near infrared spectrum quantifies
Double integrated modelling approach of analysis.
Background technology
Near-infrared spectrum technique have analyze speed is fast, simple to operate, the green advantage such as lossless and the agricultural that is widely used in,
The every field such as chemical industry, pharmacy, food, environmental protection.But the Near-infrared Spectral Absorption of complex sample is weaker, peak overlap is tight
, the interference informations such as background, noise be present, and easily influenceed by measuring condition, sample state etc., it is therefore desirable to by chemistry in weight
Meterological could carry out quantitative analysis to complex sample.When carrying out quantitative analysis to unknown sample using Chemical Measurement, prediction
As a result quality depends on the quality of model.Therefore the model for establishing high quality is in the important research of Chemical Measurement
Hold.
The conventional modeling method of Chemical Measurement has multiple linear regression, partial least-squares regression method, artificial neural network
Deng.These traditional modeling methods are predicted unknown sample only with an optimal model, sample number is less or serious interference
When precision of prediction and stability tend not to satisfactory, therefore developed integrated moulding technology.Integrated moulding technology be exactly from
Multiple training subsets are produced in same training set, then establish multiple submodels using these training subsets is predicted respectively,
And integrate multiple prediction results by certain mode, obtain a final prediction result.
Most of integrated modelling approach all carry out integrated such as bagging ELM (Bian Xihui, Li Shujuan, Tan from sample direction
Small credit, Wang Jiangjiang, Wang Zhiguo, Liu Weiguo, Chen Zongpeng, Wang Chen, extreme learning machine integrated modelling approach, 2015, Chinese patent,
ZL201510466504), boosting PLS (Bian Xihui, Liu Wei, Wang Qiunan, Tan little Yao, Guo Yugao, a kind of near infrared spectrum
Multi-model Modeling Method, Chinese invention patent, 2017, ZL201310537968.3) etc., or integrated from variable direction
(C.Tan, Qin, M.Li, Subspace regression ensemble method based on variable
Clustering for near-infrared spectroscopic calibration, Analytical Letters,
1009,42：1693-1710).If integrated from sample and variable twocouese simultaneously, multifarious son can be further produced
Model, improve model prediction accuracy.
The content of the invention
The purpose of the present invention is the problem of being directed to traditional modeling method precision of prediction and low stability, to propose a kind of near-infrared
Double integrated modelling approach (Monte Carlo-glowworm swarm algorithm-PLS, MC-FA-PLS) of quantitative spectrochemical analysis, specifically
Flow is Near-Infrared Spectra for Quantitative Analysis as shown in figure 1, to improve precision of prediction and stability to unknown sample quantitative analysis
Provide a kind of new method.
The present invention chooses a number of sample as training subset by the use of Monte Carlo sampling method (MC), using firefly
Algorithm (FA) selected part variable from the training subset as sample-variable subset, establishes partial least square model (PLS).
Successive ignition establishes some models.Final prediction result is obtained to the result progress simple average of all models.
To realize that technical scheme provided by the present invention comprises the following steps：
(1) sample of certain amount is collected, gathers the near infrared spectrum of sample, target components are determined with conventional method
Content；Training set and forecast set are splitted data into using certain packet mode；
(2) sample for choosing certain amount from training set by the use of Monte Carlo sampling method is used as training subset；
(3) glowworm swarm algorithm arameter optimization, the selected part variable conduct from training subset using glowworm swarm algorithm are carried out
Sample-variable subset；
(4) offset minimum binary factor number is determined, establishes partial least square model；
Step (2)-(4) are repeated several times, establish T model；
(5) final prediction result is obtained to the result progress simple average of all models.
The determination method of offset minimum binary factor number is：Factor number takes 25 from 1 respectively, at intervals of 1, is counted using formula (1)
Calculate the cross validation root-mean-square error (RMSECV) under different factor numbers.Examine to obtain most using Monte Carlo Cross-Validation and F
Good factor number.
M is total number of samples,For the predicted value of i-th of sample, yiFor the actual value of i-th of sample.
In the present invention, the determination method of best band number is：Training set spectrum is divided into 5~30 wave band numbers, wave band
At intervals of 5, a series of predicted root mean square errors (RMSEP) under different hop counts are calculated respectively.RMSEP values are corresponding when minimum
Hop count is best band number.
In the present invention, the determination method of glowworm swarm algorithm optimum population number is：Population number n span is 10~60,
At intervals of 10, a series of RMSEP values under different population number are calculated respectively.RMSEP values when minimum corresponding population number for most
Good population number.
In the present invention, the determination method of glowworm swarm algorithm suitable environment absorbance is：Environment absorbance γ span
For 0.1~1.2, at intervals of 0.1, a series of RMSEP values under varying environment absorbance are calculated respectively.When RMSEP values are minimum
Corresponding environment absorbance is suitable environment absorbance.
In the present invention, the determination method of glowworm swarm algorithm best constant is：The span of constant α is 0.1~1, interval
For 0.1, a series of RMSEP values under different constants are calculated respectively.RMSEP values when minimum corresponding constant be best constant.
In the present invention, the determination method of the number of the T submodel is：Give a sufficiently large submodel number
Value, the training subset sample number of fixed each data set is the 50% of gross sample number, is calculated respectively under different submodel numbers
A series of RMSEP values, when RMSEP values are almost unchanged, corresponding submodel number is required number T.
In the present invention, the sample number purpose of training subset determines that method is：Stator Number of Models, respectively from training set
Choose its total number of samples 10%~100% is used as training subset, at intervals of 5%, calculates respectively under different training subsets
A series of RMSEP values, corresponding sample number during RMSEP minimums, as required training subset sample number.
It is an advantage of the invention that：The modeling method can effectively improve the predictive ability of model, improve the prediction essence of model
Degree and stability.The inventive method is applied to the quantitative analysis of the complex samples such as oil, food, Chinese medicine.
Brief description of the drawings
Fig. 1 is MC-FA-PLS algorithm flow chart
Fig. 2 is the RMSEP values of ternary mixture data with the variation diagram of wave band number
Fig. 3 is the RMSEP values of ternary mixture data with population number n variation diagram
Fig. 4 is the RMSEP values of ternary mixture data with environment absorbance γ variation diagram
Fig. 5 is the RMSEP values of ternary mixture data with the variation diagram of constant α
Fig. 6 is the RMSEP values of ternary mixture data with the variation diagram of submodel number
Fig. 7 is the RMSEP values of ternary mixture data with the variation diagram of training subset size
Fig. 8 is the wavelength points distribution map of 80 submodels selection of ternary mixture data
Fig. 9 is the average value of predicted value and the pass of actual value that ternary mixture data run 20 times using MC-FA-PLS
System's figure
Figure 10 is the RMSEP values of fuel data with the variation diagram of submodel number
Figure 11 is the RMSEP values of fuel data with the variation diagram of training subset size
Figure 12 is the wavelength points distribution map of 80 submodels selection of fuel data
Figure 13 is the average value of predicted value and the graph of a relation of actual value that fuel data runs 20 times using MC-FA-PLS
Embodiment
To be best understood from the present invention, the present invention will be described in further detail with reference to the following examples, but of the invention
Claimed scope is not limited to the scope represented by embodiment.
Embodiment 1：
The present embodiment is analyzed using near infrared spectrum data, in the ternary mixture sample of ethanol, water and isopropanol
The measure of isopropanol content, is comprised the following steps that：
(1) the ternary mixture sample of 95 ethanol, water and isopropanol is collected, gathers the near infrared spectrum of ternary mixture
Data, spectral region 850-1049nm, 1nm is spaced, comprising 200 data points, spectrum is measured using HP-8453 spectrophotometrics
It is fixed.Using the description of the division on website to data set, 65 samples are used as training set, and 30 samples are used as forecast set；
(2) sample for choosing certain amount from training set by the use of Monte Carlo sampling method is used as training subset；
(3) glowworm swarm algorithm arameter optimization, the selected part variable conduct from training subset using glowworm swarm algorithm are carried out
Sample-variable subset；
(4) offset minimum binary factor number is determined, establishes partial least square model；
Step (2)-(4) are repeated several times, establish T model；
(5) final prediction result is obtained by carrying out simple average to the result of all models.
The determination method of offset minimum binary factor number is：Factor number takes 25 from 1 respectively, at intervals of 1, is counted using formula (1)
Calculate the RMSECV under different factor numbers.Examine to obtain optimal factor number using Monte Carlo Cross-Validation and F,
M is total number of samples,For the predicted value of i-th of sample, yiFor the actual value of i-th of sample.Calculate factor number is
7。
The determination method of best band number is：Training set spectrum is divided into 5~30 wave band numbers, wave band divides at intervals of 5
A series of RMSEP under different hop counts are not calculated.RMSEP values when minimum corresponding wave band number be best band number.The implementation
For RMSEP values as the change of wave band number is as shown in Fig. 2 when wave band number is 10, RMSEP values are minimum in example, therefore the present embodiment
Best band number is 10.
The determination method of glowworm swarm algorithm optimum population number is：Population number n span is 10~60, at intervals of 10,
A series of RMSEP values under different population number are calculated respectively.RMSEP values when minimum corresponding population number be optimum population number.
RMSEP values are as the change of population number is as shown in figure 3, when population number is 20 in the embodiment, and RMSEP values are minimum, therefore this reality
The optimum population number for applying example is 20.
The determination method of glowworm swarm algorithm suitable environment absorbance is：Environment absorbance γ span be 0.1~
1.2, at intervals of 0.1, a series of RMSEP values under varying environment absorbance are calculated respectively.RMSEP values corresponding ring when minimum
Border absorbance is suitable environment absorbance.RMSEP values are as the change of environment absorbance is as shown in figure 4, work as ring in the embodiment
When border absorbance is 0.7, RMSEP values are minimum, therefore the suitable environment absorbance of the present embodiment is 0.7.
The determination method of glowworm swarm algorithm best constant is：The span of constant α is 0.1~1, at intervals of 0.1, is distinguished
Calculate a series of RMSEP values under different constants.RMSEP values when minimum corresponding constant be best constant.In the embodiment
For RMSEP values as the change of constant is as shown in figure 5, when constant is 0.3, RMSEP values are minimum, the former the present embodiment of institute it is optimal normal
Number is 0.3.
The determination method of the number of the T submodel is：The training subset sample number of fixed each data set is gross sample
The 50% of product number, 100 sub- model number values are given, calculate a series of RMSEP values under different submodel numbers respectively,
The RMSEP values submodel number that corresponding submodel number is established for needed for when almost unchanged.In the embodiment RMSEP values with
After the change of submodel number is as shown in fig. 6, group Number of Models is 80, RMSEP values are almost unchanged, i.e., the son of required foundation
Number of Models is 80.
The sample number purpose of training subset determines that method is：Stator Number of Models 80, chooses it from training set respectively
Training subset is used as by the 10%~100% of sample number, at intervals of 5%, is calculated respectively a series of under different training subsets
RMSEP values, corresponding sample number during RMSEP minimums, as required training subset sample number.RMSEP values in the embodiment
As the change of training subset sample percentage is as shown in fig. 7, when training subset sample number reaches the 85% of training total number of samples
When, RMSEP values are minimum, therefore choose 85% that training subset sample number is gross sample number.
Fig. 8 shows the wavelength points distribution map of 80 submodel selections of the embodiment.It can be seen that all submodules
The wavelength points of type selection are largely identical, but the wavelength points of each submodel selection are again variant, ensure that the more of submodel
Sample.MC-FA-PLS models to forecast set prediction coefficient correlation be 0.9922, and PLS models to forecast set prediction
Coefficient correlation is 0.9919.It can show that MC-FA-PLS has more preferable precision of prediction by comparing.To FA-PLS and MC-FA-
PLS reruns 20 times respectively, and the standard deviation that FA-PLS standard deviation is 0.017, MC-FA-PLS is 0.012, passes through ratio
It can relatively show that MC-FA-PLS stability is better than FA-PLS.Fig. 9, which is shown, repeats 20 computings pair to MC-FA-PLS
The relation of the mean predicted value and actual value of prediction and prediction, as can be seen from the figure MC-FA-PLS prediction accuracy is high, surely
It is qualitative good.
Embodiment 2
The present embodiment is analyzed using near infrared spectrum data, to the measure of fuel oil sample rate, is comprised the following steps that：
1) 263 fuel oil samples are collected, gather the near infrared spectrum data of fuel oil sample.Spectral region 750-1550nm,
Sampling interval 2nm, comprising 400 data points, according to U.S.'s test and materials association (American Society of
Testing and Materials, ASTM) standard determine respectively, download network address：http：//
software.eigenvector.com/Data/SWRI/index.html.Using on website to data set division description,
142 samples are used as training set, and 121 samples are used as forecast set；
(2) sample for choosing certain amount from training set by the use of Monte Carlo sampling method is used as training subset；
(3) glowworm swarm algorithm arameter optimization, the selected part variable conduct from training subset using glowworm swarm algorithm are carried out
Sample-variable subset；
(4) offset minimum binary factor number is determined, establishes partial least square model；
Step (2)-(4) are repeated several times, establish T model；
(5) final prediction result is obtained by carrying out simple average to the result of all models.
The determination method of offset minimum binary factor number is：Factor number takes 25 from 1 respectively, at intervals of 1, is counted using formula (1)
Calculate the RMSECV under different factor numbers.Examine to obtain optimum factor number using Monte Carlo Cross-Validation and F,
M is total number of samples,For the predicted value of i-th of sample, yiFor the actual value of i-th of sample.Calculate factor number is
9。
The determination method of best band number is：Training set spectrum is divided into 5~30 wave band numbers, wave band divides at intervals of 5
A series of RMSEP under different-waveband number are not calculated.RMSEP values when minimum corresponding wave band number be best band number.The reality
Apply and RMSEP is observed in example with the change of wave band number, when wave band number is 20, RMSEP values are minimum, therefore the optimal ripple of the present embodiment
Hop count is 20.
The determination method of glowworm swarm algorithm optimum population number is：Population number n span is 10~60, at intervals of 10,
A series of RMSEP values under different population number are calculated respectively.RMSEP values when minimum corresponding population number be optimum population number.
RMSEP values are observed in the embodiment with the change of population number, when population number is 40, RMSEP values minimum, therefore the present embodiment
Optimum population number is 40.
The determination method of glowworm swarm algorithm suitable environment absorbance is：Environment absorbance γ span be 0.1~
1.2, at intervals of 0.1, a series of RMSEP values under varying environment absorbance are calculated respectively.RMSEP values corresponding ring when minimum
Border absorbance is suitable environment absorbance.RMSEP values are observed in the embodiment with the change of environment absorbance, when environment extinction
Spend for 0.9 when, RMSEP values are minimum, therefore the suitable environment absorbance of the present embodiment is 0.9.
The determination method of glowworm swarm algorithm best constant is：The span of constant α is 0.1~1, at intervals of 0.1, is distinguished
Calculate a series of RMSEP values under different constants.RMSEP values when minimum corresponding constant be best constant.In the embodiment
RMSEP values are observed with the change of constant, when constant is 0.5, RMSEP values are minimum, therefore the best constant of the present embodiment is
0.5。
The determination method of the number of the T submodel is：The training subset sample number of fixed each data set is gross sample
The 50% of product number, 100 sub- model number values are given, calculate a series of RMSEP values under different submodel numbers respectively,
The RMSEP values submodel number that corresponding submodel number is established for needed for when almost unchanged.In the embodiment RMSEP values with
The change of submodel number is as shown in Figure 10, and after group Number of Models is 80, RMSEP values are almost unchanged, i.e., required foundation
Submodel number is 80.
The sample number purpose of training subset determines that method is：Stator Number of Models 80, chooses it from training set respectively
Training subset is used as by the 10%~100% of sample number, at intervals of 5%, is calculated respectively a series of under different training subsets
RMSEP values, corresponding sample number during RMSEP minimums, as required training subset sample number.RMSEP values in the embodiment
As the change of training subset sample percentage is as shown in figure 11, when training subset sample number reaches the 80% of training total number of samples
When, RMSEP values are minimum, therefore choose 80% that training subset sample number is gross sample number.
Figure 12 shows the wavelength points distribution map of 80 submodel selections of the embodiment.It can be seen that all sons
The wavelength points of model selection are largely identical, but the wavelength points of each submodel selection are again variant, ensure that submodel
Diversity.MC-FA-PLS models to forecast set prediction coefficient correlation be 0.9856, and PLS models to forecast set predict
Coefficient correlation be 0.9687.It can show that MC-FA-PLS has more preferable precision of prediction by comparing.To FA-PLS and MC-
FA-PLS reruns 20 times respectively, and the standard deviation that FA-PLS standard deviation is 0.0019, MC-FA-PLS is 0.0017,
It can show that MC-FA-PLS stability is better than FA-PLS by comparing.Figure 13 is shown to be repeated 20 times to MC-FA-PLS
Computing is to the mean predicted value predicted and predicted and the relation of actual value, and as can be seen from the figure MC-FA-PLS prediction is accurate
Degree is high, and stability is good.
Claims (3)
1. double integrated modelling approach of a kind of Near-Infrared Spectra for Quantitative Analysis, it is characterised in that concretely comprise the following steps：
(1) sample of certain amount is collected, gathers the near infrared spectrum of sample, is contained using conventional method measure target components
Amount；Certain packet mode is taken to split data into training set and forecast set；
(2) sample for choosing certain amount from training set by the use of Monte Carlo sampling method is used as training subset；
(3) carry out glowworm swarm algorithm arameter optimization, using glowworm swarm algorithm from training subset selected part variable as sample-
Variable subset；
(4) offset minimum binary factor number is determined, establishes partial least square model；
Step (2)-(4) are repeated several times, establish T model；
(5) final prediction result is obtained to the result progress simple average of all models.
A kind of 2. double integrated modelling approach of Near-Infrared Spectra for Quantitative Analysis according to claim 1, it is characterised in that：Institute
The determination method for stating the number of T submodel is：Give a sufficiently large submodel number value, the instruction of fixed each data set
Practice 50% that subset sample number is gross sample number, calculate a series of RMSEP values under different submodel numbers respectively, when
RMSEP values when almost unchanged corresponding submodel number be required submodel number T.
A kind of 3. double integrated modelling approach of Near-Infrared Spectra for Quantitative Analysis according to claim 1, it is characterised in that：Institute
The sample number purpose for stating training subset determines that method is：Stator Number of Models, by the 10%~100% of sample number, at intervals of
5%, a series of RMSEP values under different training subsets are calculated respectively, and sample number corresponding to RMSEP minimums is required instruction
Practice subset sample number.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710569312.8A CN107356556A (en) | 2017-07-10 | 2017-07-10 | A kind of double integrated modelling approach of Near-Infrared Spectra for Quantitative Analysis |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710569312.8A CN107356556A (en) | 2017-07-10 | 2017-07-10 | A kind of double integrated modelling approach of Near-Infrared Spectra for Quantitative Analysis |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107356556A true CN107356556A (en) | 2017-11-17 |
Family
ID=60291958
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710569312.8A Pending CN107356556A (en) | 2017-07-10 | 2017-07-10 | A kind of double integrated modelling approach of Near-Infrared Spectra for Quantitative Analysis |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107356556A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2021036546A1 (en) * | 2019-08-29 | 2021-03-04 | 山东科技大学 | Near-infrared quantitative analysis model construction method based on biased estimation |
| CN112924413A (en) * | 2021-01-27 | 2021-06-08 | 四川长虹电器股份有限公司 | Method for predicting vinasse components |
| CN113607683A (en) * | 2021-08-09 | 2021-11-05 | 天津九光科技发展有限责任公司 | Automatic modeling method for near infrared spectrum quantitative analysis |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102735642A (en) * | 2012-07-12 | 2012-10-17 | 河北大学 | Method for quickly and losslessly identifying virgin olive oil and olive-residue oil |
| CN103528990A (en) * | 2013-10-31 | 2014-01-22 | 天津工业大学 | Method for establishing multiple models of near infrared spectrums |
| CN105117525A (en) * | 2015-07-31 | 2015-12-02 | 天津工业大学 | Bagging extreme learning machine integrated modeling method |
| CN105203498A (en) * | 2015-09-11 | 2015-12-30 | 天津工业大学 | Near infrared spectrum variable selection method based on LASSO |
| CN106529008A (en) * | 2016-11-01 | 2017-03-22 | 天津工业大学 | Double-integration partial least square modeling method based on Monte Carlo and LASSO |
-
2017
- 2017-07-10 CN CN201710569312.8A patent/CN107356556A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102735642A (en) * | 2012-07-12 | 2012-10-17 | 河北大学 | Method for quickly and losslessly identifying virgin olive oil and olive-residue oil |
| CN103528990A (en) * | 2013-10-31 | 2014-01-22 | 天津工业大学 | Method for establishing multiple models of near infrared spectrums |
| CN105117525A (en) * | 2015-07-31 | 2015-12-02 | 天津工业大学 | Bagging extreme learning machine integrated modeling method |
| CN105203498A (en) * | 2015-09-11 | 2015-12-30 | 天津工业大学 | Near infrared spectrum variable selection method based on LASSO |
| CN106529008A (en) * | 2016-11-01 | 2017-03-22 | 天津工业大学 | Double-integration partial least square modeling method based on Monte Carlo and LASSO |
Non-Patent Citations (1)
| Title |
|---|
| 刘泽蒙 等: ""基于离散萤火虫算法的近红外波长优选方法研究"", 《光谱学与光谱分析》 * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2021036546A1 (en) * | 2019-08-29 | 2021-03-04 | 山东科技大学 | Near-infrared quantitative analysis model construction method based on biased estimation |
| CN112924413A (en) * | 2021-01-27 | 2021-06-08 | 四川长虹电器股份有限公司 | Method for predicting vinasse components |
| CN113607683A (en) * | 2021-08-09 | 2021-11-05 | 天津九光科技发展有限责任公司 | Automatic modeling method for near infrared spectrum quantitative analysis |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN103528990B (en) | A kind of multi-model Modeling Method of near infrared spectrum | |
| CN104062257B (en) | A kind of based on the method for general flavone content near infrared ray solution | |
| CN105388123B (en) | A kind of method by near infrared spectrum prediction oil property | |
| CN107817223A (en) | The construction method of quick nondestructive real-time estimate oil property model and its application | |
| CN106529008B (en) | A kind of double integrated offset minimum binary modeling methods based on Monte Carlo and LASSO | |
| CN101413885A (en) | Near-infrared spectrum method for rapidly quantifying honey quality | |
| CN105300923A (en) | Correction method of measuring point free temperature compensation model during process of online application of near infrared spectrum analyzer | |
| Huang | Chemometric methods in analytical spectroscopy technology | |
| CN102590129B (en) | Method for detecting content of amino acid in peanuts by near infrared method | |
| CN101825567A (en) | Screening method for near infrared spectrum wavelength and Raman spectrum wavelength | |
| CN106650926B (en) | A kind of steady boosting extreme learning machine integrated modelling approach | |
| CN104020127A (en) | Method for rapidly measuring inorganic element in tobacco by near infrared spectrum | |
| CN109324013A (en) | A method of it is quickly analyzed using Gaussian process regression model building oil property near-infrared | |
| CN108956583A (en) | Characteristic spectral line automatic selecting method for laser induced breakdown spectroscopy analysis | |
| Li et al. | Nondestructive detection of frying times for soybean oil by NIR-spectroscopy technology with Adaboost-SVM (RBF) | |
| CN107356556A (en) | A kind of double integrated modelling approach of Near-Infrared Spectra for Quantitative Analysis | |
| CN106950192A (en) | A kind of method of Contents of Main Components quick detection in vegetable protein beverage based on near-infrared spectral analysis technology | |
| Wang et al. | Onsite nutritional diagnosis of tea plants using micro near-infrared spectrometer coupled with chemometrics | |
| CN115993344A (en) | Quality monitoring and analyzing system and method for near infrared spectrum analyzer | |
| CN102128805A (en) | Method and device for near infrared spectrum wavelength selection and quick quantitative analysis of fruit | |
| CN106018331A (en) | Stability evaluation method of multi-channel spectrum system and pretreatment optimization method | |
| CN103134770A (en) | Method for eliminating influence on infrared spectrum detection of soil total nitrogen content from moisture | |
| CN102854151B (en) | Chemometrics method for classifying sample sets in spectrum analysis | |
| CN107271389B (en) | A kind of spectral signature variable fast matching method based on index extreme value | |
| Yang et al. | Feasibility of terahertz spectroscopy for hybrid purity verification of rice seeds |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| CB03 | Change of inventor or designer information |
Inventor after: Bian Xihui Inventor after: Wu Xinyan Inventor after: Qiu Jianming Inventor after: Liu Peng Inventor after: Tan Xiaoyao Inventor after: Li Ming Inventor before: Bian Xihui Inventor before: Qiu Jianming Inventor before: Liu Peng Inventor before: Tan Xiaoyao Inventor before: Li Ming |
|
| CB03 | Change of inventor or designer information | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20171117 |
|
| RJ01 | Rejection of invention patent application after publication |